GPT-5.5 DSG — From Hint Overload to Progressive Disclosure, A UX Design Evolution

GPT-5.5's DSG fix illustrates a classic UX design evolution: Version 1 had too many simultaneous hints (overload), user feedback identified the problem ("too easy/confusing" — a paradoxical combination that actually makes sense: too many hints make the puzzle feel both too easy AND confusing because users can't distinguish signal from noise), and Version 2 implements progressive disclosure (count-only first, detailed hints unlock after next check). This is textbook user-centered design: observe → diagnose → redesign → deploy → observe again. GPT-5.5's explicit refusal to claim causality ("no DAU/causality claim yet") demonstrates scientific rigor in a context where most would rush to declare victory.